3.1.1. Total density

We calculated a basic community characteristic, the total density of individuals, to see if patterns could be detected in the study area. The same calculation as for Chapter 1 has been performed for the considered stations.

When we considered the data without a distinction by station, the global total density is 3606 individuals.grab^-1 for shallow stations and 7309 individuals.grab^-1 for deep stations (bay-scale is 10,915 individuals.grab^-1 when pooled).

3.1.2. Total biomass

We calculated a basic community characteristic, the total biomass of individuals, to see if patterns could be detected in the study area.

When we considered the data without a distinction by station, the global total density is 191.1581 gWM.grab^-1 for shallow stations and 715.056 gWM.grab^-1 for deep stations (bay-scale is 906.2141 gWM.grab^-1 when pooled).

3.1.3. W statistic

This indicator is based on abundance and biomass ranked values, as presented by Warwick (1986) and Clarke (1990). In addition to Abundance-Biomass Curves (see above), it allows to present a disturbed state thanks to a certain structure of the community.

The W statistic is continuous between -1 and 1, and is calculated using this equation:

\[ W = \frac{\sum_{i = 1}^{S}(B_{i} - A_{i})}{50(S - 1)} \]

• \(B_{i}\) is the biomass of a species
• \(A_{i}\) is the abundance of a species
• \(S\) is the specific richness
• \(i\) is a species rank

When we considered the data without a distinction by station, the global W statistic is 0.1343 for shallow stations and 0.1064 for deep stations (bay-scale is 0.1035 when pooled).

3.2.1. Specific richness

We calculated a basic community characteristic, the specific richness, to see if patterns could be detected in the study area. The same calculation as for Chapter 1 has been performed for the considered stations.

ASSUMPTION: A higher richness indicates a high status without perturbation.

When we considered the data without a distinction by station, the global specific richness is 65 for shallow stations and 117 for deep stations (bay-scale is 132 when pooled).

3.2.2. Shannon index

We calculated basic a community characteristic, the Shannon index, to see if patterns could be detected in the study area. The same calculation as for Chapter 1 has been performed for the considered stations.

ASSUMPTION: A higher index indicates a high status without perturbation.

When we considered the data without a distinction by station, the global Shannon index is 2.667023 for shallow stations and 3.180947 for deep stations (bay-scale is 3.251545 when pooled).

3.2.3. Margalef index

We calculated a basic community characteristics, the Margalef index, to see if patterns could be detected in the study area.

ASSUMPTION: A higher index indicates a high status without perturbation. (To check)

When we considered the data without a distinction by station, the global Margalex index is 7.81407 for shallow stations and 13.03831 for deep stations (bay-scale is 14.09715 when pooled).

3.2.4. Simpson index

We calculated a basic community characteristics, the Simpson index, to see if patterns could be detected in the study area.

When we considered the data without a distinction by station, the global Simpson index is 0.8784821 for shallow stations and 0.9230898 for deep stations (bay-scale is 0.9266169 when pooled).

3.2.5. Pielou evenness

We calculated a basic community characteristics, the Pielou evenness, to see if patterns could be detected in the study area. The same calculation as for Chapter 1 has been performed for the considered stations.

When we considered the data without a distinction by station, the global Pielou index is 0.6389016 for shallow stations and 0.6679611 for deep stations (bay-scale is 0.6659178 when pooled).

3.2.6. Taxonomic diversity

We calculated a basic community characteristic, the taxonomic diversity, to see if patterns could be detected in the study area. The same calculations as for Chapter 1 have been performed for the considered stations.

When we considered the data without a distinction by station, the global taxonomic diversity is 68.47852 for shallow stations and 74.80171 for deep stations (bay-scale is 74.16541 when pooled).

3.2.7. Functional diversity

We studied functional diversity based on five biological traits and 26 modalities:

• body composition: non calcified tissue, calcareous, calcareous calcium carbonate, calcareous amorphous calcium carbonate, calcareous aragonite, calcareous calcite, calcareous high magnesium calcite, chitinous
• body size: small, medium, large
• food diet: filter, surface deposit, subsurface deposit, predator, scavenger, grazer, parasite
• mobility: sessile, limited, mobile
• lifestyle: fixed, tubicolous, burrower, crawler, swimmer

Species were assigned a value for each modality using a scale varying from 0 (absence of the modality) to 1 (presence), where the sum of the values for every modality of a trait equals 1. This allowed to calculate functional richness, evenness and divergence according to Laliberté & Legendre (2010).

For some reason, R is not able to calculate a global value…

3.3.1. AZTI Marine Biotic Index (AMBI)

AMBI (initially named the biotic coefficient) is an ecological index that is used to detect a perturbation in an ecosystem based on the composition of the communities (Borja et al., 2000). This perturbation is linked with an organic matter increase, according to Pearson and Rosenberg (1978) model.

To compute the index, species are classed into five groups in relation to their tolerance to this perturbation:

• group I (GI): vulnerable species
• group II (GII): indifferent species
• group III (GIII): tolerant species
• group IV (GIV): first-order opportunistics
• group V (GV): second-order opportunistics

These groups are based on expert opinion on the physiology of species and experimental studies, but the attribution of a species to a group can be somewhat arbitrary (e.g. based on related phyla information) so it needs to be interpretated carefully.

AMBI is continuous between 0 to 6 (7 when the habitat is azoic, thus considered extremely disturbed), and is calculated using this equation in a dedicated software:

\[ AMBI = \frac{\sum_{i}^{GI-V} w_{i} . P_{i}}{100} \]

• \(P_{i}\) is the proportion of each group (percentage of the total density of species)
• \(w_{i}\) is the weighting parameter of each group (respectively 0, 1.5, 3, 4.5 and 6)
• \(i\) is the ecological group

ASSUMPTION: Sensitive species are only present in pristine ecosystems, while the dominance of opportunists indicates a perturbed state.

Here is the classification of the sampled taxa in ecological groups, adapted from Borja et al. (2000):

When we considered the data without a distinction by station, the global AMBI index is 1.571 for shallow stations and 1.529 for deep stations (bay-scale is 1.543 when pooled).

3.3.2. Multivariate AMBI (M-AMBI)

M-AMBI is a complementary method to AMBI that is based on a multivariate ordination (factorial analysis and discriminant analysis) of the stations using the species richness, the Shannon index and the AMBI index (Muxika et al. 2007). This index needs to explicitely define reference conditions, corresponding to ‘bad’ and ‘high’ conditions.

These values are called ‘references’, but this vision can be biaised. We have calculated them using the 5 % and 95 % percentiles of the distribution shallow and deep stations. This is a recommendation by Nicolas Desroy, so that we do not detect a modification of quality status when there is a small perturbation (see work by Pearson & Rosenberg and the Intermediate Disturbance Hypothesis).

This calculation provided :

• S = 2.5, H = 0.35, AMBI = 2.88 for ‘bad’ conditions at shallow stations
• S = 16.25, H = 2.01, AMBI = 0.72 for ‘high’ conditions at shallow stations
• S = 7, H = 1.14, AMBI = 2.43 for ‘bad’ conditions at deep stations
• S = 21.95, H = 2.54, AMBI = 0.78 for ‘high’ conditions at deep stations

M-AMBI is continuous between 0 and 1, and is calculated using a dedicated software.

ASSUMPTION: A high richness, high diversity and low AMBI index indicate a high status without perturbation.

As we do not have historical data or other reference systems to produce relevant reference conditions, it is difficult to calculate a global M-AMBI here.

No clear tendancy can be discovered here, apart from the fact that the overall status seems to be ‘High’. Several hypothesises can explain this result:

• the M-AMBI index describes reality well, so that overall perturbation from organic matter is low
• there is a bias in the index due to the species classification in groups, originally suited for European ecosystems
• the assumptions for the reference values are not correct
• the configuration of the bay makes the perturbation small relative to the water volume and bathymetric condition

Further work is needed to determine the individual responses of somes species, along with the use of different methods to understand other perturbations and cumulative impacts.

3.3.3. BENTIX

BENTIX is an index based on the same theory as the AMBI, where species are placed in groups based on their tolerance to perturbation (Simboura & Zenetos, 2002). Here also, this perturbation is principally linked to organic matter increase, but two groups only are present:

• GS: species that are sensitive or indifferent to a perturbation (~ AMBI groups I and II)
• GT: species that are tolerant to a perturbation and opportunists (~ AMBI groups III to V)

BENTIX is continuous between 2 and 6 (0 when the habitat is azoic, thus considered extremely disturbed), and is calculated using this equation:

\[ BENTIX = \frac{(6 . P_{GS}) + (2 . P_{GT})}{100} \]

• \(P_{GS}\) is the proportion of sensitive species (percentage of the total density of species)
• \(P_{GT}\) is the proportion of tolerant species (percentage of the total density of species)

ASSUMPTION: Sensitive species are only present in pristine ecosystems, while the dominance of opportunists indicates a perturbed state.

Here is the classification of the sampled taxa in ecological groups, adapted from Simboura & Zenetos (2002):

When we considered the data without a distinction by station, the global BENTIX index is 5.194676 for shallow stations and 5.224518 for deep stations (bay-scale is 5.213659 when pooled).

3.3.4. Benthic opportunistic polychaete/amphipod ratio (BOPA)

BOPA is an index that uses a relative abundance ratio of species in a community to infer a state of perturbation. Ratios with many species have been tested, and opportunistic polychaetes and amphipods have been selected to be the most pertinent (originally to detect effects of an oil-spill on soft-bottom communities, e.g. from the Sea Empress or the Amoco Cadiz). It has been updated from its original form in 2000.

BOPA is continuous between 0 and \(log_{10}(2)\) (~ 0.3), and is calculated using this equation:

\[ BOPA = \left( \frac{f_{P}}{f_{A} + 1} + 1 \right) \]

• \(f_{P}\) is the relative frequency of opportunistic polychaetes (abundance / total density)
• \(f_{A}\) is the relative frequency of amphipods (abundance / total density)

We considered ecological groups GIII to GV for polychaetes and GI for amphipods (without Jassa genera).

ASSUMPTION: Dominance of amphipods characterizes pristine ecosystems, while a dominance of opportunistic polychaetes indicates a perturbed state.

These are the polychaetes and amphipods present in our species list (including the confidence score used during group classification).

When we considered the data without a distinction by station, the global BOPA index is 0.002157566 for shallow stations and 0.004086439 for deep stations (bay-scale is 0.003443646 when pooled).